Adaptive multiple access interference suppression

• US 6,700,923 B1
• Filed: 12/29/1999
• Issued: 03/02/2004
• Est. Priority Date: 01/04/1999
• Status: Expired due to Fees

• Current Assignee: Board of Regents of the University of Texas System (University of Texas System)
• Original Assignee: Board of Regents of the University of Texas System (University of Texas System)
• Inventors: Dowling, Eric M., Jani, Umesh G., Wang, Zifei, Golden, Richard M.
• Primary Examiner(s): Chin, Stephen
• Assistant Examiner(s): WILLIAMS, LAWRENCE B
• Application Number:
  US09/474,653
• Time in Patent Office: 1,525 Days
• Field of Search: 375/140, 375/148, 375/208, 375/342, 370/201, 370/291, 370/342
• US Class Current: 375/148
• CPC Class Codes: H04B 1/7105 Joint detection techniques, e.g. linear detectors

Abstract

An apparatus and method for spread spectrum multiuser communication receivers is disclosed that includes, an adaptive filter adapted to minimize a block error function, wherein the filter tap weights are updated according to a non-linear optimization algorithm adapted for block processing of time-varying data.

Claims

1. A spread spectrum receiver comprising a block former coupled to receive samples of a spread spectrum waveform, said samples being sampled at a specified chip rate, said block former also being coupled to a memory and operative to store a block of data therein;

a gradient calculation module coupled to said memory and to a tap-weight vector calculation module, said gradient calculation module is operative to receive said data block from said memory and a tap weight vector from said tap weight vector calculation module, said gradient calculation module further being operative to compute a gradient vector with respect to an error function with respect to said tap weight vector;

a search direction calculation module coupled to receive a gradient vector from said gradient calculation module and coupled to receive a previous search direction via a feedback path, said search direction calculation module operative to compute a search direction as a function of said previous search direction and the two most recently computed gradient vectors;

a stepsize calculation module coupled to said memory, and to said tap weight vector calculation module, said stepsize calculation unit operative to compute a step size according to a prespecified criterion;

a tap-weight vector calculation module coupled to receive input from said search direction calculation module and said step size calculation module, and a previous tap weight vector via a feedback path, said tap weight vector calculation module operative to compute a subsequent search direction as a function of said previous tap weight vector, said step size and said previous tap weight vector to produce an updating rule; and

an output module operative to couple a linear combination of said tap weight vector with a selected set of said samples of said spread spectrum waveform to produce a block of output symbol values.

2. The method of claim 1, wherein said updating rule comprises the modified block Shanno constant modulus algorithm.

3. The method of claim 1, wherein said updating rule minimizes a constant modulus objective function.

4. The method of claim 1, wherein said updating rule does not require c step size parameters to be set a priori based on signal and channel power considerations.

5. The method of claim 1, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using block processing.

6. The method of claim 1, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using non-Toeplitz block processing.

7. The method of claim 1, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a step size that satisfies restrictions at each update.

8. The method of claim 1, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using multiple updates per block.

9. The method of claim 1, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm that provides for a gradient descent at each update.

10. The method of claim 1, wherein said updating rule optimizes a nonlinear objective function using a block difference objective function.

11. The method of claim 1, wherein said method comprises CDMA transmissions.

12. The method of claim 1, wherein said method comprises TDMA transmissions.

13. The method of claim 1, wherein said method is implemented in cellular telephone transmissions.

14. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm adapted for time-varying environments using blocks of data.

15. The method of claim 14, wherein said updating rule is the modified block Shanno constant modulus algorithm.

16. The method of claim 14, wherein said updating rule minimizes a constant modulus using an objective function.

17. The method of claim 14, wherein said updating rule does not require a set of step size parameters to be set a priori based on signal and channel power considerations.

18. The method of claim 14, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using block processing.

19. The method of claim 14, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a step size that satisfies restrictions at each update.

20. The method of claim 14, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using multiple updates per block.

21. The method of claim 14, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a block gradient descent at each update.

22. The method of claim 14, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a block difference objective function.

23. The method of claim 14, wherein said updating rule optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a non-Toeplitz block difference objective function.

24. The method of claim 14, wherein said method comprises CDMA transmissions.

25. The method of claim 14, wherein said method comprises TDMA transmissions.

26. A CDMA receiver structure comprising:

an adaptive filter with an input coupled to an analog to digital converter and an output coupled to a decision device, said adaptive filter comprising a filter tap weight vector, wherein said filter tap weight vector is initialized with at least a portion of a spreading code associated with a user;

a block-nonlinear filter update processor coupled to receive a block of data from said analog to digital converter and said adaptive filter tap weigh vector, said block-nonlinear filter update processor having an output coupled to provide a new set of tap weight values back to said adaptive filter, said block-nonlinear filter update processor operative to perform one or more iterations of a nonlinear optimization algorithm on said block of data.

27. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule is a modified block Shanno constant modulus algorithm.

28. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm, adapted for time-varying environments using blocks of data, and minimizes a constant modulus using an objective function.

29. A method of extracting a data sequence from a spread spectrum waveforrn, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm, adapted for time-varying environments using blocks of data, and does not require a set of step size parameters to be set a priori based on signal and channel power considerations.

30. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm, adapted for time-varying environments using blocks of data, and optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using block processing.

31. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm, adapted for time-varying environments using blocks of data, and optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a step size that satisfies restrictions at each update.

32. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm, adapted for time-varying environments using blocks of data, and optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using multiple updates per block.

33. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm, adapted for time-varying environments using blocks of data, and optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a block gradient descent at each update.

34. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm, adapted for time-varying environments using blocks of data, and optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a block difference objective function.

35. A method of extracting a data sequence from a spread spectrum waveform, the data sequence being associated with a specified user, the method comprising steps of:

receiving an input spread spectrum signal and sampling said signal at a chip rate;

initializing an adaptive filter with a despreading code, said despreading code associated with said specified user;

filtering said spread spectrum signal by said adaptive filter to produce an output, said filtering occurring at said chip rate, and said filtering producing outputs at a symbol rate;

comparing said output signal to at least one threshold level and deciding on the value of a recovered data symbol based on said comparison to produce a recovered output symbol; and

updating a set of filter coefficients using an updating rule, said filter coefficients being used in said step of filtering, wherein said updating rule implements a nonlinear optimization algorithm, adapted for time-varying environments using blocks of data, and optimizes a nonlinear objective function using a nonlinear optimization algorithm adapted for adaptive processing using a non-Toeplitz block difference objective function.